Detection of Diethyl Phthalate in Perfumes by
Extractive Electrospray Ionization Mass
Spectrometry
Konstantin Chingin,
†
Huanwen Chen,*
,‡
Gerardo Gamez,
†
Liang Zhu,
†
and Renato Zenobi*
,†
Department of Chemistry and Applied Biosciences, ETH Zu¨rich CH-8093 Zu¨rich Switzerland, and Applied Chemistry
Department, East China Institute of Technology, Fuzhou, China
Recent findings suggest that long-term exposure to diethyl
phthalate (DEP), one of the widely used phthalate esters,
can lead to serious health problems. Most perfumes
contain non-negligible amounts of DEP. Rapid and sensi-
tive detection of DEP in perfumes is thus of increasing
importance. A novel procedure based on extractive elec-
trospray ionization mass spectrometry (EESI-MS) has
been developed for fast detection and identification of
DEP in perfumes without the need for any sample pre-
treatment. The limit of determination for DEP in perfume
was less than 100 ppb using tandem mass spectrometry
on a commercial quadrupole time-of-flight mass spec-
trometer. The dynamic range of this method was about 4
orders of magnitude. A single sample analysis was com-
pleted within a few seconds, providing a rapid way to
obtain semiquantitative information on the DEP content
in perfumes. This study shows that both volatile and
nonvolatile analytes (e.g., amino acids) in liquids can be
directly sampled by neutral desorption, providing a con-
venient way for high-throughput screening of target com-
pounds using EESI-MS.
Diethyl phthalate (DEP) is a plasticizer widely used in many
industrial products, including tools, automotive parts, tooth-
brushes, food packaging, cosmetics and insecticides.
1,2
DEP is
also widely used in the perfume industry as a vehicle for
fragrances and as an alcohol denaturizing agent. Until recently it
was believed that cosmetics containing phthalates pose no risks
to human health or the environment. Thus, currently relatively
high amounts of DEP are used in most cosmetics such as
perfumes. DEP concentrations higher than a few percent occur
in many products available on the market. In a recent survey by
Greenpeace, DEP was found in 34 out of 36 perfumes tested.
3
The highest levels of DEP were detected in Eternity by Calvin
Klein (2.2% (w/w)), Iris Blue by Melvita (1.1% (w/w)), and Le
Male by Jean-Paul Gaultier (0.99% (w/w)). DEP concentrations
in perfumes have not been regulated by existing legislation
because they were believed to have low overall toxicity.
4,5
Recent
findings, however, contest their safety.
6-16
Studies on rats have
shown that upon long-term exposure, DEP can be a reason for
reproductive failure.
6,7,9,12,13
Reduced neurotransmitter activity due
to water contamination with DEP was observed in adult male
fishes.
10,15
Recently, DEP was also shown to be a possible
promoter of ocular damage in animals.
16
Besides animal experi-
ments, developmental and reproductive toxicity of DEP is sug-
gested by recent human studies.
17,18
In 2005, the European Union
banned the use of six phthalates including DEP in children’s
products.
19
When applied to skin, DEP rapidly penetrates it and
becomes widely distributed around the body following each
exposure.
20
This renders a perfume that contains DEP potentially
hazardous.
* To whom correspondence should be addressed. Prof. Dr. Renato Zenobi,
Chemistry Department and Applied Biosciences, ETH Zu¨rich, CH-8093 Zu¨rich,
Huanwen Chen, Applied Chemistry Department, East China Institute of Technol-
†
ETH Zurich.
‡
East China Institute of Technology.
(1) Schettler, T. Int. J. Androl. 2006, 29, 134–139
.
(2) Wormuth, M.; Scheringer, M.; Vollenweider, M.; Hungerbuhler, K. Risk
Anal. 2006, 26, 803–824
.
(3) Perivier, H. An Investigation of Chemicals in Perfumes; Greenpeace
International, 2005;pp1-16.
(4) Api, A. M. Food Chem. Toxicol. 2001, 39, 97–108
.
(5) Gray, L. E.; Ostby, J.; Furr, J.; Price, M.; Veeramachaneni, D. N. R.; Parks,
L. Toxicol. Sci. 2000, 58, 350–365
.
(6) Pereira, C.; Mapuskar, K.; Rao, C. V. Regul. Toxicol. Pharmacol. 2006, 45,
169–177
.
(7) Mapuskar, K.; Pereira, C.; Rao, C. V. Pestic. Biochem. Phys. 2007, 87, 156–
163
.
(8) Rozati, R.; Reddy, P. P.; Reddanna, P.; Mujtaba, R. Fertil. Steril. 2002, 78,
1187–1194
.
(9) Pereira, C.; Mapuskar, K.; Rao, C. V. Environ. Toxicol. Pharmacol. 2007,
23, 319–327
.
(10) Barse, A. V.; Chakrabarti, T.; Ghosh, T. K.; Pal, A. K.; Jadhao, S. B. Pestic.
Biochem. Phys. 2007, 88, 36–42
.
(11) Pereira, C.; Mapuskar, K.; Rao, C. V. Pestic. Biochem. Phys. 2007, 88, 156–
166
.
(12) Pereira, C.; Mapuskar, K.; Rao, C. V. Pestic. Biochem. Phys. 2008, 90, 52–
57
.
(13) Fujii, S.; Yabe, K.; Furukawa, M.; Hirata, M.; Kiguchi, M.; Ikka, T. J. Toxicol.
Sci. 2005, 30 Spec No., 97–116
.
(14) Latini, G.; Del Vecchio, A.; Massaro, M.; Verrotti, A.; De Felice, C. Toxicology
2006, 226, 90–98
.
(15) Ghorpade, N.; Mehta, V.; Khare, M.; Sinkar, P.; Krishnan, S.; Rao, C. V.
Ecotoxicol. Environ. Saf. 2002, 53, 255–258
.
(16) Askari, S. N.; Zaidi, M.; Ahmad, N. Turk. J. Med. Sci. 2006, 36 (4), 231–
234
.
(17) Swan, S. H.; Main, K. M.; Liu, F.; Stewart, S. L.; Kruse, R. L.; Calafat, A. M.;
Mao, C. S.; Redmon, J. B.; Ternand, C. L.; Sullivan, S.; Teague, J. L. Environ.
Health Perspect. 2005, 113, 1056–1061
.
(18) Colon, I.; Caro, D.; Bourdony, C. J.; Rosario, O. Environ. Health Perspect.
2000, 108, 895–900
.
(19) http://www.environmentcalifornia.org/environmental-health/stop-toxic-
toys.
(20) http://www.inchem.org/documents/cicads/cicads/cicad52.htm.
Anal. Chem. 2009, 81, 123–129
10.1021/ac801572d CCC: $40.75 2009 American Chemical Society
123Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
Published on Web 12/08/2008
Many techniques including chromatography,
21-23
optical
spectroscopy,
24,25
ion mobility spectrometry,
26
and mass spec-
trometry
27-32
have been used for detection of DEP in various
samples. Because of the complexity of the real samples, a sample
cleanup procedure was usually required for most techniques in
previous studies. Chromatographic techniques such as GC and
HPLC are conventionally used to identify phthalate esters in
cosmetic products.
33,34
A single HPLC/GC run typically takes
approximate 30 min. In addition, sample pretreatment may be
required, especially for sensitive detection (e.g., extraction, pre-
concentration). However, when one has to process a large number
of samples (e.g., in quality control laboratories), high throughput
may become of major importance. An appropriate analytical tool
for practical sample analysis must meet the demanding require-
ments for high throughput, high sensitivity, and specificity since
the concentration of DEP as well as the sample composition varies
over a wide range.
To facilitate high-throughput mass spectrometric analysis,
techniques such as desorption electrospray ionization (DESI),
35-37
desorption atmospheric pressure chemical ionization (DAPCI),
38-41
direct analysis in real time (DART),
36,41-44
atmospheric-pressure
solids analysis probe (ASAP),
45,46
and thermal desorption atmo-
spheric pressure chemical ionization (TD-APCI)
47,48
have been
used for fast detection of analytes on solid surfaces. Besides some
common advantages of these ambient ionization techniques, each
one has unique features for specific analytical applications. Liquid
samples can be analyzed by these techniques but require sample
pretreatment. For example, dilute urine dried on a paper surface
49
can be examined by DESI. Similar to DESI, other techniques such
as DART, ASAP, and TD-APCI analyze liquid samples indirectly.
Usually, deposition of the sample on a solid surface and solvent
volatilization is employed as a sample preparation step. In DART,
TD-APCI as well as ASAP, the solvent evaporation can be
shortened to 2-3 min by heating the sample surface to a high
temperature (e.g., 250-450 °C). However, such a high tempera-
ture results in fast degradation of sensitive compounds, and thus
the mass spectrum of a sample can be significantly changed. This
renders data interpretation difficult, especially for analysis of heat
sensitive samples such as perfumes.
Liquids, gases, suspensions, and aerosol samples can be
directly analyzed by extractive electrospray ionization (EESI)
36,50-62
mass spectrometry without any sample pretreatment. With the
use of a neutral desorption (ND) device,
36,53,54
analytes such as
metabolites, active drug components, explosives, and chemical
pollutants can be liberated from virtually any type of surface for
subsequent EESI analysis. Recently we demonstrated that ND-
EESI-MS enables rapid classification of perfumes without any
sample pretreatment.
63
Perfumes are complex liquid samples,
which are composed of fragrant essential oils, aroma compounds,
fixatives, and alcohol matrixes. Note that many ingredients in
perfumes are heat sensitive and degrade quickly when heated to
(21) Lambropoulou, D. A.; Konstantinou, I. K.; Albanis, T. A. J. Chromatogr., A
2007, 1152, 70–96
.
(22) Kato, K.; Silva, M. J.; Needham, L. L.; Calafat, A. M. Anal. Chem. 2005,
77, 2985–2991
.
(23) Kato, K.; Silva, M. J.; Needham, L. L.; Calafat, A. M. Anal. Chem. 2006,
78, 6651–6655
.
(24) Du, Q.; Shen, L.; Xiu, L.; Jerz, G.; Winterhalter, P. Food Addit. Contam.
2006, 23, 552–555
.
(25) Steiner, H.; Jakusch, M.; Kraft, M.; Karlowatz, M.; Baumann, T.; Niessner,
R.; Konz, W.; Brandenburg, A.; Michel, K.; Boussard-Pledel, C.; Bureau,
B.; Lucas, J.; Reichlin, Y.; Katzir, A.; Fleischmann, N.; Staubmann, K.;
Allabashi, R.; Bayona, J. M.; Mizaikoff, B. Appl. Spectrosc. 2003, 57, 607–
613
.
(26) Baumbach, J. I.; Eiceman, G. A. Appl. Spectrosc. 1999, 53, 338A–355A
.
(27) Ezerskis, Z.; Morkunas, V.; Suman, M.; Simoneau, C. Anal. Chim. Acta
2007, 604, 29–38
.
(28) Martin, A. N.; Farquar, G. R.; Frank, M.; Gard, E. E.; Fergenson, D. P.
Anal. Chem. 2007, 79, 6368–6375
.
(29) Silva, M. J.; Samandar, E.; Preau, J. L.; Reidy, J. A.; Needham, L. L.; Calafat,
A. M. J. Chromatogr., B 2007, 860, 106–112
.
(30) Cao, X. L. J. Chromatogr., A 2008, 1178, 231–238
.
(31) Carrillo, J. D.; Martinez, M. P.; Tena, M. T. J. Chromatogr., A 2008, 1181,
125–130
.
(32) Nilsson, C.; Viberg, P.; Spegel, P.; Jornten-Karlsson, M.; Petersson, P.;
Nilsson, S. Anal. Chem. 2006, 78, 6088–6095
.
(33) Shen, H. Y.; Jiang, H. L.; Mao, H. L.; Pan, G.; Zhou, L.; Cao, Y. F. J. Sep.
Sci. 2007, 30, 48–54
.
(34) De Orsi, D.; Gagliardi, L.; Porra, R.; Berri, S.; Chimenti, P.; Granese, A.;
Carpani, I.; Tonelli, D. Anal. Chim. Acta 2006, 555, 238–241
.
(35) Chen, H. W.; Talaty, N. N.; Takats, Z.; Cooks, R. G. Anal. Chem. 2005,
77, 6915–6927
.
(36) Venter, A.; Nefliu, M.; Cooks, R. G. TrAC, Trends Anal. Chem. 2008, 27,
284–290
.
(37) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306,
471–473
.
(38) Chen, H. W.; Liang, H. Z.; Ding, J. H.; Lai, J. H.; Huan, Y. F.; Qiao, X. L. J.
Agric. Food Chem. 2007, 55, 10093–10100
.
(39) Chen, H. W.; Lai, J. H.; Zhou, Y. F.; Huan, Y. F.; Li, J. Q.; Zhang, X.; Wang,
Z. C.; Luo, M. B. Chin. J. Anal. Chem. 2007, 35, 1233–1240
.
(40) Chen, H. W.; Zheng, J.; Zhang, X.; Luo, M. B.; Wang, Z. C.; Qiao, X. L. J.
Mass Spectrom. 2007, 42, 1045–1056
.
(41) Williams, J. P.; Patel, V. J.; Holland, R.; Scrivens, J. H. Rapid Commun.
Mass Spectrom. 2006, 20, 1447–1456
.
(42) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297–
2302
.
(43) Moffat, A. C.; Cody, R. B.; Jee, R. D.; O’Neil, A. J. J. Pharm. Pharmacol.
2007, 59, A26–A26
.
(44) Kpegba, K.; Spadaro, T.; Cody, R. B.; Nesnas, N.; Olson, J. A. Anal. Chem.
2007, 79, 5479–5483
.
(45) McEwen, C. N.; McKay, R. G.; Larsen, B. S. Anal. Chem. 2005, 77, 7826–
7831
.
(46) McEwen, C.; Gutteridge, S. J. Am. Soc. Mass Spectrom. 2007, 18, 1274–
1278
.
(47) Sleeman, R.; Burton, I. F. A.; Carter, J. F.; Roberts, D. J. Analyst 1999,
124, 103–108
.
(48) Ebejer, K. A.; Brereton, R. G.; Carter, J. F.; Ollerton, S. L.; Sleeman, R.
Rapid Commun. Mass Spectrom. 2005, 19, 2137–2143
.
(49) Chen, H. W.; Pan, Z. Z.; Talaty, N.; Raftery, D.; Cooks, R. G. Rapid Commun.
Mass Spectrom. 2006, 20, 1577–1584
.
(50) Jackson, A. U.; Werner, S. R.; Talaty, N.; Song, Y.; Campbell, K.; Cooks,
R. G.; Morgan, J. A. Anal. Biochem. 2008, 375, 272–281
.
(51) Chen, H. W.; Zenobi, R. Chimia 2007, 61, 843–843
.
(52) Chen, H. W.; Touboul, D.; Jecklin, M. C.; Zheng, J.; Luo, M. B.; Zenobi,
R. N. Eur. J. Mass Spectrom. 2007, 13, 273–279
.
(53) Chen, H.; Yang, S.; Wortmann, A.; Zenobi, R. Angew. Chem., Int. Ed. 2007,
46, 7591–7594
.
(54) Chen, H. W.; Wortmann, A.; Zenobi, R. J. Mass Spectrom. 2007, 42, 1123–
1135
.
(55) Zhou, Z. Q.; Jin, M.; Ding, J. H.; Zhou, Y. M.; Zheng, J.; Chen, H. W.
Metabolomics 2007, 3, 101–104
.
(56) Chen, H. W.; Sun, Y. P.; Wortmann, A.; Gu, H. W.; Zenobi, R. Anal. Chem.
2007, 79, 1447–1455
.
(57) Chen, H. W.; Wortmann, A.; Zhang, W. H.; Zenobi, R. Angew. Chem., Int.
Ed. 2007, 46, 580–583
.
(58) Gu, H. W.; Chen, H. W.; Pan, Z. Z.; Jackson, A. U.; Talaty, N.; Xi, B. W.;
Kissinger, C.; Duda, C.; Mann, D.; Raftery, D.; Cooks, R. G. Anal. Chem.
2007, 79, 89–97
.
(59) Chen, H. W.; Venter, A.; Cooks, R. G. Chem. Commun. 2006, 2042–2044
.
(60) Martinez-Lozano, P.; de la Mora, J. F. Int. J. Mass Spectrom. 2007, 265,
68–72
.
(61) Zhu, L.; Gamez, G.; Chen, H. W.; Huang, H. X.; Chingin, K.; Zenobi, R.
Rapid Commun. Mass Spectrom. 2008, 22, 2993–2998
.
(62) Zhu, L.; Gamez, G.; Chen, H. W.; Chingin, K.; Zenobi, R. Chem. Commun.
2008, in press, DOI: 10.1039/B818541G.
(63) Chingin, K.; Gamez, G.; Chen, H. W.; Zhu, L.; Zenobi, R. Rapid Commun.
Mass Spectrom. 2008, 22, 2009–2014
.
124 Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
a temperature higher than 200 °C.
64
In conventional desorption
sampling, it takes only ∼2-3 min for a perfume deposited on a
solid substrate to dry. The speed is not a serious problem in most
applications. However, it is necessary to improve the throughput
and to simplify the data interpretation when a specific analyte
(such as DEP) in a large number of complex samples (such as
perfume products) must be rapidly screened. Herein we extend
EESI-MS to rapid detection and identification of DEP directly from
liquid perfume samples, without any sample pretreatment.
EXPERIMENTAL SECTION
The key part of the experimental setup is schematically shown
in Figure 1. A pipet tip (Gilson, France) filled with 1 µLofa
perfume was exposed to a gentle stream of nitrogen gas (100∼400
L/h) coming out from a Teflon tube witha7mminnerdiameter.
The pipet tip was 2 cm away from the sampling cone of the ESI
source (Z-spray, Micromass, U.K.). The sample was desorbed and
then transported from the pipet tip by the neutral nitrogen gas
flow. Perfume droplets thus produced were ionized by the EESI
source using a solvent mixture (methanol/water/acetic acid 40%/
40%/20%), infused at a flow rate of 5 µL/min. The experiments
were run in the positive ion detection mode on a commercial
electrospray ionization (ESI) quadrupole time-of-flight (Q-TOF)
mass spectrometer (QTOF Ultima, Micromass, Manchester,
U.K.). Briefly, the capillary voltage was 3 kV and the cone voltage
was 40 V. Other parameters were default values of the instrument.
No further optimization was performed.
For DEP identification, tandem mass spectrometry was used.
Ions of m/z 223 corresponding to protonated DEP ions were
isolated in an rf hexapole and then subjected to collisions with
buffer gas molecules in the collision cell. Fragments as well as
nondissociated parent ions were then detected by the TOF
analyzer. Collision energy was set to 10 units on the QTOF
software. This allowed observing the parent ion and two major
fragments at m/z 177 and m/z 149. These three peaks were finally
used to identify DEP in perfume samples.
The Mass Lynx 4.0 software (Waters, Manchester, U.K.) was
used for the QTOF-MS experiments. A detailed procedure for
background subtraction in a QTOF instrument has been described
elsewhere.
65
The mass spectra were typically accumulated for
about 5-10 s, the single scan time being 0.5 s.
Eighteen fragrances by different brands were examined on
DEP content: “Weekend” by Burberry, “Relaxing fragrance” by
Shiseido, “Be delicious” by DKNY, “Beautiful” by Estee Lauder,
“Hugo XY” by Hugo Boss, “le Male” by Jean Paul Gaultier, “ETH
Zurich 150” by Givadaun, “Bright Crystal” by Versace, “Option”
by Nova, “CK One” by Calvin Klein, “Miss Dior” and “Midnight
Poison” by Christian Dior, “Clinique Happy for Men” and “Clinique
Happy Heart” by Clinique, “Opium”, “Opium Shanghai”, and
“Opium Impereale” by Yves Saint Laurent, and “White Musk” (Eau
de Toilette) by The Body Shop.
DEP (>99.5% purity) was purchased from Fluka (Switzerland).
Chemicals such as methanol and acetic acid were bought from
Fluka (Switzerland) with the highest purity grades available. The
water used was deionized water, available in house at ETH. To
minimize the background signal of DEP, plastic material was
avoided completely for handling the reagents. The only exception
was a plastic pipet tip (D10, Gilson, France) used to deliver
perfume samples to neutral desorption EESI. These pipet tips were
made from pure, translucent polypropylene. No additives are used
in their production. No background level change was observed
when glass tips were used instead, in agreement with the
presumed absence of DEP in these tips. Therefore, cheap
replaceable plastic tips were used in all the experiments, which
is also beneficial for high-throughput analysis. For quantitative
analysis, perfumes were diluted with ethanol 1000 times to
minimize matrix effects, followed by a standard addition analysis
using DEP.
Safety Recommendations. Latex powder-free gloves were
used in all the experiments in order to avoid undesirable exposure
of skin to solvents. It is recommended to use an exhaust vent
over the nebulization area in order to reduce exposure to
inhalation of fumes and prevent substantial DEP deposition in the
ion source area.
RESULTS AND DISCUSSION
Rapid Screening of DEP by EESI-MS. In our previous study
on perfumes, neutral desorption was carried out from samples
deposited on a smelling strip.
63
The analytes in the perfume were
then sampled to the EESI source by neutral desorption after the
strip had dried. However, for rapid analysis of liquid samples,
especially in cases where only a particular compound is of interest,
neutral desorption from a liquid hanging droplet is a more
straightforward sampling method, and facilitates high-throughput
analysis by eliminating the step to deposit and dry the sample on
a surface. Also, sensitive detection from a substrate surface (e.g.,
paper) can be complicated because of adhesion of DEP as well
as other components to the surface. Stronger nitrogen gas flows
would then be required to release analytes from the surface. This,
however, would result in a decreased residence time of the
desorbed neutrals inside the ionizing plume, yielding lower overall
sensitivity.
With the use of a gentle gas flow (as shown in Figure 1), the
hanging perfume droplet was nebulized and the resulting aerosol
was transferred into an ESI plume for ionization. Figure 2 presents
EESI-MS fingerprints of three different fragrances (“le Male” by
Jean Paul Gaultier; “Hugo XY” by Hugo Boss, and “Natural
fragrance” by Shiseido) within the m/z 100-300 range. As can
be seen, several peaks are commonly present in all the three mass
spectral patterns. These signals typically correspond to the
generally used essential fragrance compounds in the perfume
industry, such as limonene (m/z 137) and citronellol (m/z 157);
(64) Burr, C. The Emperor of Scent: A Story of Perfume, Obsession, and the Last
Mystery of the Senses; Random House: New York, 2003.
(65) Chen, H.; Zenobi, R. Nat. Protoc. 2008, 3, 1467–1475
.
Figure 1. Schematics of the experimental setup where liquids are
sampled directly by gentle neutral desorption for EESI-MS detection.
125Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
both of them showed up as protonated molecules in the EESI-
MS spectra. However, even though these compounds occur in
all the three perfume samples, their relative abundances are very
different (e.g., citronellol and limonene). This is obvious from the
different relative peak intensities for the corresponding com-
pounds in the mass spectra. Some signals, on the other hand,
occur only in particular fragrances (e.g., m/z 246 in “le Male”).
The signal at m/z 223 is abundant in the EESI-MS fingerprint for
“le Male” perfume, noticeable in “Hugo XY” and was not detected
in “Natural fragrance”. In order to judge whether this signal comes
from protonated DEP species or some other compound with the
same molecular weight, its MS/MS spectrum should be compared
to the reference MS/MS spectrum of pure DEP. Figure 3 shows
an EESI-MS/MS spectrum of authentic DEP (1 ppm in ethanol).
One can see protonated DEP ions at m/z 223 as well as the first
and the second fragment at m/z 177 and m/z 149, respectively.
Figure 3 shows the structures of the characteristic fragments.
66
The fragment at m/z 177 started to appear at lower CID energies
than that at m/z 149, indicating that the pathway for ethanol
cleavage was favored. Identical MS/MS spectral patterns of ions
(m/z 223) were observed in perfume samples as well. These data
confirmed the successful detection of DEP in perfumes. In order
to achieve high throughput analysis, the CID energy was
optimized to see both characteristic fragments simultaneously
(CID energy was 10 units in the QTOF software).
EESI-MS/MS spectra of 18 perfume samples were obtained
under the same conditions to identify the DEP content. DEP was
detected in 13 out of the 18 samples (detected in “Weekend”,
“Beautiful”, “Hugo XY”, “le Male”, “ETH Zurich 150”, “Bright
Crystal”, “Option”, “CK One”, “Miss Dior”, “Clinique Happy for
Men”, “Opium”, “Opium Shanghai”, “Opium Impereale”; not
detected in “Natural fragrance”, “White Musk”, “Midnight Poison”,
“Be delicious”, and “Clinique Happy Heart”). For these samples
MS/MS spectral patterns of m/z 223 ions were identical to that
of authentic DEP. DEP was found in abundance in perfume
products for both men and women. The highest detected signal
was from ‘”le Male” by Jean Paul Gaultier. According to the
Greenpeace Survey, this fragrance holds about 1% DEP (w/w).
3
Most products designed for men contained a measurable percent-
age of DEP. Also, our observations indicate a huge variation of
DEP content in perfumes, supporting the recent results from the
Greenpeace survey.
3
Recent findings suggest that DEP is more
toxic to males by causing reproductive failure,
6,7,9,12,13
thus it is
of greater concern for men to avoid using perfume products
containing high amounts of DEP.
It is worth noting that for the samples that we examined the
chemical and detector noise level was always negligible compared
to the detected signals from DEP in perfumes. This means that
we did not meet any borderline case when the DEP signal was
around the detection limit, there was either an easily seen signal
or no signal at all. Typical ion current profiles of the two major
DEP fragments for “Miss Dior” perfume are presented in Figure
4. The plateau region in the profile corresponds to the time when
the perfume was being sampled for EESI analysis. Notably, the
ion current responded rapidly to the presence of perfume samples.
For example, the 90% signal raise time was less than1sforthe
fragment of m/z 177, and after a measurement time of ∼0.5 min,
the signal dropped down to the noise level in about 0.8 s when
the pipet with a perfume sample was moved away. Note that the
full scan time was 0.5 s for this experiment, i.e., the signal
responded within two scans. In none of the experiments performed
was any sample carryover observed. Therefore, the analysis speed
is largely dependent on the measurement time, which can be
shortened to less than 1 s using a fast scan mode. For the current
setup, the speed is essentially limited by the sample delivery. In
our experiments, the perfume samples were delivered manually
using a pipet; it took about 1∼3 s for each sample loading. Clearly,
this simple method enables high-throughput detection of DEP.
(66) McLafferty, F. W. Interpretation of Mass Spectra, 3rd ed.; University Science
Books: Mill Valley, CA, 1980.
Figure 2. Chemical fingerprints of three famous fragrances, recorded
in the positive ion detection mode by EESI-MS. Top, “le Male” by
Jean Paul Gaultier; middle, “Hugo XY” by Hugo Boss; bottom, “Natural
fragrance” by Shiseido.
Figure 3. MS/MS spectrum of DEP recorded using ND-EESI-MS.
Figure 4. Ion “chromatograms” of DEP fragments and total ion
current (TIC) detected by EESI-MS/MS from a “Miss Dior” perfume.
126 Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
Taking advantage of the unique design of EESI, ion/molecule
reactions can be easily implemented in the ionization process. It
was found that DEP formed ionic sodium adducts (m/z 245) rather
than protonated molecules when a diluted sodium chloride
solution (1 ppm) was electrosprayed in the EESI source. This
information can enhance the specificity of DEP detection from
complex perfume samples. When the sodium content was de-
creased down to the low parts per billion range, both of the sodium
adducts and the protonated molecules of DEP were simultaneously
observed in the EESI-MS spectra. Cationization of DEP can
provide a better sensitivity, probably due to higher affinity of DEP
to sodium ions compared to protons. This is also supported by
the CID experiments of DEP sodium adducts, which gave no
fragments with the possible highest energy for collisions in our
QTOF instrument, showing the stability of the DEP sodium
adducts. Thus, a reliable identification without resorting to tandem
MS can be done from a high precision mass measurement
experiments of the parent ions by using FTICR
67
or Orbitrap
68
analyzers, both of which feature very high mass accuracy.
Also, it is well-known that the CID fragment of DEP at m/z
149 is characteristic for other phthalates as well.
66
Therefore,
surveying all parent ions of this fragment could be informative of
other closely related phthalates present in a perfume sample. This
can be easily achieved by neutral-loss scan available in a number
of tandem mass spectrometers, e.g., triple-stage quadrupoles.
69
Therefore, this method can be potentially extended to fast
screening of any phthalic acid in perfume products.
Possible Sampling Mechanism. In this study a neutral gas
flow was used in order to sample a drop of liquid to EESI analysis.
This approach is analogous to the technique introduced earlier
for sampling analytes deposited on a solid surface known as
“neutral desorption”.
36,53,54,65
For consistency with these earlier
studies, we also use the term “neutral desorption” here when
referring to the sampling method. The mechanism of analyte
liberation from the bulk liquid, however, may be different. Since
DEP is a semivolatile compound, we believe that two processes
contribute to the sample desorption-transportation process. The
first is a simple evaporation assisted by the neutral gas flow, which
delivers gas-phase DEP molecules into the surrounding air. The
second mechanism is a process based on nebulization of the
sample liquid. In this process, small neutral droplets are produced
first with the assistance of the desorption gas beam, by disruption
of the liquid surface. These droplets are carried by the neutral
desorption gas, moving toward to the EESI source in the air. They
form an aerosol containing both droplets with dissolved DEP and
DEP vapor produced by normal evaporation. The aerosol is
transported to the EESI region for extractive electrospray ioniza-
tion, and finally partially solvated DEP species are ionized. An
important difference between these two mechanisms is that
droplet formation does not require the analyte to be volatile. Figure
5 shows an EESI mass spectrum of arginine dissolved in water
(100 µM). The spectrum was recorded under the same experi-
mental conditions that were used for DEP detection. Arginine is
a nonvolatile amino acid. No signal could be detected when the
neutral desorption gas was off. This demonstrates that even
nonvolatile samples can be transported from the liquid by neutral
desorption. Further, we carried out a test experiment in which
neutral desorption was performed on rhodamine 6G (1 mM in
ethanol) under the same conditions used for perfume analyses.
Figure 6 shows aerosol droplets of the rhodamine 6G solution
after deposition on a piece of white paper. As rhodamine 6G is a
typical nonvolatile compound, the essential factor contributing to
its extraction from a large droplet on the pipet tip was aerosol
generation induced by the neutral gas flow.
Sensitivity and Dynamic Range. In order to determine the
sensitivity of the detection, a sample that did not produce any
DEP signal (“White Musk” Eau de Toilette) was spiked with DEP
(Note that “White Musk” Eau de Toilette is a fragrance different
from “White Musk” Eau de Parfum examined by Greenpeace
3
).
In the MS/MS experiments, the characteristic fragments (i.e., m/z
177, 149) of DEP could be simultaneously detected starting from
a concentration of 100 ppb (w/w) DEP in the perfume. The total
amount of DEP was quite low since only 1 µL of solution was
used. The detection limit was about 10 times lower when DEP
was diluted in a pure alcohol, e.g., methanol or ethanol, rather
than a complex perfume matrix. This finding suggests that even
though perfumes largely consist of alcohols, the neutral desorption
efficiency depends on the viscosity of a particular perfume sample.
We also noticed that maximum yields of the first and the second
(67) Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. Rev.
1998, 17, 1–35
.
(68) Hu, Q. Z.; Noll, R. J.; Li, H. Y.; Makarov, A.; Hardman, M.; Cooks, R. G. J.
Mass Spectrom. 2005, 40, 430–443
.
(69) deHoffmann, E. J. Mass Spectrom. 1996, 31, 129–137
.
Figure 5. A mass spectrum of arginine in aqueous solution recorded
using ND-EESI.
Figure 6. Aerosols produced by neutral desorption from a droplet
of ethanol solution of rhodamine 6G. The fine droplets were recorded
using a piece of white paper, which was placed parallel to the gas
flow direction. The distance from the pipet tip to the paper surface
was 5 mm. The color is due to deposition of rhodamine molecules
on the paper surface. The extension of the plume was about 5 cm.
The sample volume in the pipet was 2 µL. The neutral desorption
gas flow was 400 L/h.
127Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
fragments occur at different collisional energy values. For ex-
ample, the fragment at m/z 177 could be detected alone with a
much higher intensity than when both fragments at m/z 177 and
149 showed up (as shown in Figure 3). In order to increase the
sensitivity of the method, one can detect specific fragments at
different CID energy values. The detection limit can thus be lower
than 100 ppb in cases where the fragment m/z 177 is known to
be exclusively derived from DEP, and then only this fragment
can be used for quantification.
Another factor limiting the sensitivity rather than the meth-
odology itself was chemical noise in the mass spectrometer.
Phthalates are widely used in the production of PVC and other
plastics as plasticizers. Such additives are not chemically bound
to the polymer, i.e., they are continuously released into the
ambient air or leached out.
70
This makes phthalic acid esters
ubiquitous contaminants. For example, they can be extracted from
the ESI source plastic tubing by the solvents used.
71
Also, the
solvent itself can be contaminated by phthalates because of their
leaching from plastic containers.
72
For most MS instruments,
gases such as sheath gas are supplied through plastic tubes. As
plastic materials commonly contain phthalates, DEP can be
desorbed by the nitrogen gas flowing through the tubing and then
delivered to the ionizing region. It was impossible to replace all
the tubes used in our instrument, i.e., further attempts to eliminate
the chemical background due to DEP would require a complete
redesign of the ESI source. Fortunately, as typical DEP signals
from perfume samples are relatively high, it was not necessary to
get an absolutely clean background signal for screening of DEP
in perfume samples.
Quantitative Determination of DEP. Since perfumes are
complex samples, it is necessary to exclude false positive signal
by using tandem mass spectrometry to utilize the major fragments
for quantitative and qualitative analysis. Figure 7 shows the
dependence of the MS/MS signal intensity of the fragment at m/z
177 on the DEP concentration in ethanol. The response curve
was an exponential, with a linearity coefficient R
2
) 0.9989 in a
doubly logarithmic representation, showing a dynamic range
of 4 orders of magnitude. The figure suggests a detection limit
far below 100 ppb; however, as discussed above, it is difficult
to obtain a clean background. The quantitative signal response
is also dependent on the matrixes of the samples. For most
commercial perfume products, the major matrix is ethanol.
However, as mentioned above, the same amount of DEP spiked
into a DEP-free perfume matrix and into pure ethanol gave
quite different signal levels. In order to overcome this matrix
effect, standard addition was used for quantifying DEP in
perfume samples.
73
Moreover, perfume samples were diluted
in pure ethanol to make sure that the signal response would
be linear, which is not the case for high DEP concentrations,
as shown in Figure 7
. Dilution was continued until the S/N ratio
for the DEP signal was below 100, which represented a good
compromise between sufficient sensitivity and elimination of
matrix effects. The diluted solution was then spiked with known
amounts of DEP (standard additions). Six independent measure-
ments were made for each standard addition to record a response
curve. The latter was extrapolated to give the unknown concentra-
tion of DEP in a perfume sample as shown in Figure 8 for “le
Male”. The same procedure was used for quantitative analysis of
DEP in “CK One”. Concentrations obtained were 1.1 ± 0.3%
(w/w) for DEP in “le Male” and 0.28 ± 0.09% for “CK One”. The
same fragrances were studied in the Greenpeace survey,
3
and the
DEP content reported was 0.99% and 0.11% for “le Male” and “CK
One” accordingly. This reasonably good agreement suggests that
our method can also be used for rapid semiquantitative analysis
of the DEP content in perfumes.
CONCLUSIONS
Neutral desorption of liquid perfume samples using a gentle
gas beam for rapid extractive electrospray ionization was dem-
(70) Heudorf, U.; Mersch-Sundermann, V.; Angerer, E. Int. J. Hyg. Environ.
Health 2007, 210, 623–634
.
(71) Jenke, D. R.; Story, J.; Lalani, R. Int. J. Pharm. 2006, 315, 75–92
.
(72) Bosnir, J.; Puntaric, D.; Galic, A.; Skes, I.; Dijanic, T.; Klaric, M.; Grgic,
M.; Curkovic, M.; Smit, Z. Food Technol. Biotechnol. 2007, 45, 91–95
.
(73) Harris, D. Quantitative Chemical Analysis, 6th ed.; W. H. Freeman: New
York, 2003.
Figure 7. Dependence of the MS/MS signal intensity at m/z 177 on
DEP concentration in ethanol.
Figure 8. Quantification of the DEP content in “le Male” perfume
using standard addition. The sample was diluted 1000 times in ethanol
to make sure the response was linear. Signals were recorded from
the diluted sample spiked with 10, 25, and 50 ppm of DEP along
with a nonspiked sample. Concentration of DEP in the diluted sample
was extracted by extrapolating the linear fit to the x axis (213.45x +
2394.2 ) 0, x ) 11.2 ppm). The number thus obtained was multiplied
by 1000 to give the actual concentration of DEP in the nondiluted
sample. The error bars show the standard deviation of the mean value
of six measurements for each standard addition.
128 Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
onstrated. The concept of neutral desorption sampling was
extended for application to liquid samples for the first time. A
novel method was established to semiquantitatively detect the
content of DEP in various perfumes by tandem EESI-MS. DEP is
a phthalate that is potentially toxic to humans but is still widely
used in industry and in many products. Many countries are
starting to regulate DEP contents in products such as toys.
However, many perfume products contain significant amount of
DEP, which requires a sensitive and easy-to-implement method
for rapid screening of perfumes that may contain notable amounts
of DEP. With the use of the method reported here, no sample
pretreatment is required for perfume analysis, and a single sample
analysis can be completed within a few seconds. The limit of
determination for DEP in perfume was on the order of 100 ppb
with tandem mass spectrometry. This method provided a dynamic
response range about 4 orders of magnitude, providing a rapid
way to obtain semiquantitative information on DEP in bulk
perfume analyses. Furthermore, our experimental data show that
both volatile and nonvolatile analytes in complex liquid samples
can be directly sampled by neutral desorption. This sampling
method provides a high duty cycle of analysis since absolutely
no sample pretreatment is required. Because the liquid sample
can be delivered quickly using cheap disposable devices such as
pipets, this method is a convenient way for high-throughput
screening of target compounds in liquid samples. For applications
in which trace amounts of analytes need to be continuously
monitored, the classic configuration of EESI using two spray
beams
50,58,59
is probably preferable, because it enables online, real
time monitoring of complex samples with ease. However, when
high-throughput analysis is required (e.g., identification of hazard-
ous species in bulk commercial products on the market), our novel
sample delivery method becomes particularly useful.
ACKNOWLEDGMENT
This work was partly supported by a grant from NNSFC (Grant
20505003).
Received for review July 25, 2008. Accepted November
18, 2008.
AC801572D
129Analytical Chemistry, Vol. 81, No. 1, January 1, 2009
